Pros of moving IBM BPM / BAW to the cloud (CP4BA on OpenShift in a public cloud, or the SaaS offerings):
- No WebSphere cell to install and patch: operators apply upgrades; sizing by replicas; environments are namespaces.
- Elastic capacity for peaks (month-end, campaigns) and cheaper non-production environments (scale to zero at night).
- Modern security stack (IAM / OIDC, secrets, TLS by default), audited platform, DR by cluster replication.
- Access to the rest of the Cloud Pak (ADS decisions, RPA, Document Processing, BAI) on the same platform; SaaS removes operations entirely.
- Licensing by vCPU with the Cloud Pak metering, often cheaper for variable loads than PVU-based on-prem licenses.
Cons:
- Migration effort: heritage human services / coaches are not supported on containers, some Java integrations and file-system tricks stop working, IID / BPEL (Advanced) is not available - a code review before the move (question 2989).
- Network: back ends stay on-prem - latency and VPN / private link for every integration; LDAP has to be reachable or federated.
- Skills: OpenShift, operators, YAML and observability replace WebSphere console skills; troubleshooting changes (pod logs, metrics).
- Data residency / regulation: where instance data and documents live; encryption keys under your control (KMS).
- Cost predictability: cloud infrastructure + Cloud Pak licenses + egress; SaaS has the least control (upgrade windows, customisation limits).
- Lock-in is not worse than on-prem (same product), but the OpenShift requirement is real (no plain Kubernetes / Docker).
A useful decision path: (1) inventory heritage artifacts and integrations; (2) pilot one process app on CP4BA in the cloud with a VPN to one back end; (3) measure latency and cost for a month; (4) decide between CP4BA-managed-by-you and the SaaS offering based on how much customisation the pilot needed.
References